9652662

Image Processing Device and Image Processing Method

PublishedMay 16, 2017
Assigneenot available in USPTO data we have
Technical Abstract

Patent Claims
5 claims

Legal claims defining the scope of protection, as filed with the USPTO.

1

1. An image processing device comprising: a face detector which detects a face region including a face of a person from an input image signal; a determination portion which determines, in the input image signal of the face region, an addition amount of a pixel value to be added to the input image signal of the face region based on a size of a difference in a pixel value between adjacent pixels; and a signal corrector which adds the addition amount to the pixel value of the input image signal of the face region, wherein a pixel having a larger pixel value of the input image signal is defined to be convex compared to a pixel having a smaller pixel value of the input image signal in the face region, the determination portion includes: a direction detector which detects a direction of the face in the face region; a light source setting portion which sets an illumination direction of a light source virtually provided to illuminate the face based on the direction of the face; a normal line estimation portion which estimates a normal line direction of a surface of the face included in the face region based on the size of the difference in the pixel value and the definition; an addition amount calculator which calculates the addition amount based on the normal line direction estimated by the normal line estimation portion and the illumination direction of the light source set by the light source setting portion; and, a light source corrector which corrects the illumination direction of the light source set by the light source setting portion, the light source setting portion sets the illumination direction of the light source such that the face is illuminated from below of a front of the face, the normal line estimation portion estimates the normal line direction such that the normal line direction becomes closer to parallel to a plane including the face region as the difference in the pixel value becomes larger, the addition amount calculator calculates the addition amount such that the addition amount becomes smaller as an angle formed by the normal line direction and the illumination direction of the light source becomes larger, the input image signal is input for each frame, the direction detector detects the direction of the face for each frame, the light source setting portion sets a light source vector which represents the illumination direction of the light source based on the direction of the face for each frame, a frame immediately before a first frame is defined as a second frame, the light source corrector corrects the light source vector set by the light source setting portion for each frame, the addition amount calculator calculates the addition amount using the light source vector corrected by the light source corrector for each frame, and a corrected light source vector Lc(t) of the first frame corrected by the light source corrector is calculated using a light source vector L(t) of the first frame set by the light source setting portion and a corrected light source vector Lc(t−1) obtained when a light source vector L(t−1) of the second frame set by the light source setting portion is corrected by the light source corrector.

2

2. The image processing device according to claim 1 , wherein the addition amount calculator calculates the addition amount from the angle formed by the normal line direction and the illumination direction of the light source, and from a pixel value of each pixel of the input image signal.

3

3. An image processing device comprising: a face detector which detects a face region including a face of a person from an input image signal; a determination portion which determines, in the input image signal of the face region, an addition amount of a pixel value to be added to the input image signal of the face region based on a size of a difference in a pixel value between adjacent pixels; a signal corrector which adds the addition amount to the pixel value of the input image signal of the face region; a movement detector which detects, for each frame, a movement vector between adjacent frames based on the input image signal input for each frame; and a region corrector which corrects the face region detected by the face detector based on the movement vector to output a corrected face region, wherein a pixel having a larger pixel value of the input image signal is defined to be convex compared to a pixel having a smaller pixel value of the input image signal in the corrected face region, the determination portion includes: a normal line estimation portion which estimates a normal line direction of a surface of the face included in the corrected face region based on the size of the difference in the pixel value and the definition, a light source information storage which stores an illumination direction and a light intensity of a light source virtually provided to illuminate the face, and an addition amount calculator which calculates an addition amount of a pixel value to be added to the input image signal of the corrected face region based on the normal line direction estimated by the normal line estimation portion and the illumination direction and the light intensity of the light source stored in the light source information storage, a frame immediately before a first frame is defined as a second frame, the face detector detects the face region for each frame, the region corrector corrects the face region for each frame, the normal line estimation portion estimates the normal line direction such that the normal line direction becomes closer to parallel to a plane including the corrected face region as the difference in the pixel value becomes larger, the addition amount calculator calculates the addition amount such that the addition amount becomes smaller as an angle formed by the normal line direction and the illumination direction of the light source becomes larger, the signal corrector adds the addition amount to the pixel value of the input image signal of the corrected face region, and a corrected face region Rc(t) of the first frame corrected by the region corrector is calculated using a face region R(t) of the first frame detected by the face detector and a region Rv(t−1) obtained when a face region R(t−1) of the second frame detected by the face detector is moved by the movement vector.

4

4. The image processing device according to claim 3 , wherein the determination portion further includes a normal line corrector which corrects a normal line vector, which represents the normal line direction estimated by the normal line estimation portion, based on the movement vector for each frame, the addition amount calculator calculates the addition amount using the corrected normal line vector corrected by the normal line corrector for each frame, and a corrected normal line vector Nc(t) of the first frame corrected by the normal line corrector is calculated using a normal line vector N(t) of the first frame, which represents the normal line direction estimated by the normal line estimation portion, and a normal line vector Nv(t−1) obtained when a normal line vector N(t−1) of the second frame, which represents the normal line direction estimated by the normal line estimation portion, is moved by the movement vector.

5

5. An image processing device comprising: a face detector which detects a face region including a face of a person from an input image signal; a determination portion which determines, in the input image signal of the face region, an addition amount of a pixel value to be added to the input image signal of the face region based on a size of a difference in a pixel value between adjacent pixels; and a signal corrector which adds the addition amount to the pixel value of the input image signal of the face region, wherein a pixel having a larger pixel value of the input image signal is defined to be convex compared to a pixel having a smaller pixel value of the input image signal in the face region, the determination portion includes: a normal line estimation portion which estimates a normal line direction of a surface of the face included in the face region based on the size of the difference in the pixel value and the definition; a light source information storage which stores an illumination direction and a light intensity of a light source virtually provided to illuminate the face; and an addition amount calculator which calculates the addition amount based on the normal line direction estimated by the normal line estimation portion and the illumination direction and the light intensity of the light source stored in the light source information storage, the normal line estimation portion estimates the normal line direction such that the normal line direction becomes closer to parallel to a plane including the face region as the difference in the pixel value becomes larger, the addition amount calculator calculates the addition amount such that the addition amount becomes smaller as an angle formed by the normal line direction and the illumination direction of the light source becomes larger, respective pixels of the input image signal are arranged in a matrix in a first direction and in a second direction orthogonal to the first direction, the normal line estimation portion includes: a first differential processor which differentiates a pixel value of an estimation target pixel, for which the normal line direction is estimated, in the first direction to output a first differential value; a second differential processor which differentiates the pixel value of the estimation target pixel in the second direction to output a second differential value; and a spatial filter processor which smooths the first differential value to output a first smoothed value and smooths the second differential value to output a second smoothed value, and the normal line estimation portion retains a fixed value determined in advance, and generates a normal line vector, which represents the normal line direction, with the first smoothed value as a component in the first direction of the normal line vector, the second smoothed value as a component in the second direction of the normal line vector, and the fixed value as a component in a third direction of the normal line vector.

Patent Metadata

Filing Date

Unknown

Publication Date

May 16, 2017

Inventors

Tsuyoshi HIRASHIMA
Tatsumi WATANABE

Want to explore more patents?

Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.

Citation & reuse

Analysis on this page is generated by Patentable — an AI-powered patent intelligence platform. AI-generated summaries, explanations, and analysis may be reused with attribution and a visible link back to the canonical URL below. Patent abstracts and claims are USPTO public domain.

Cite as: Patentable. “IMAGE PROCESSING DEVICE AND IMAGE PROCESSING METHOD” (9652662). https://patentable.app/patents/9652662

© 2026 Patentable. All rights reserved.

Patentable is a research and drafting-assistant tool, not a law firm, and does not provide legal advice. Documents we generate are drafts for review by a licensed patent attorney.